Could computers and IT, rather than destroying the traditional media, help us understand what’s wrong with it?
While there are many problems with traditional media, the major one is probably the increasing tabloidisation of even broadsheet newspapers which sometimes seem to compete with the shock jocks.
Part of that problem is the apparent inability of much of the media – either for structural or conceptual reasons – to demonstrate any more than a limited ability to discriminate between noise and signals.
The most important insights into most areas – from science to sociology and investing in the stock market – come when people manage to discern the signal amongst the noise. Equally the most significant errors come when people either can’t hear the signal or imagine one that isn’t there. Hence, data deluges on politics, economic and investment news and even the week’s sporting events can lead to confusion or imagined significance where none exists.
A friend sent me an article recently, which I thought at first, might provide an explanation of the latter. A computer program, DISCERN, has recently thrown some light on how schizophrenia works. According to an article in the IEEE Spectrum (11 June 2011) researchers at the University of Texas Austin have been modelling schizophrenia and came up with a new hypothesis on how it works. According to the Spectrum article by Morgen E. Peck, DISCERN (short for distributed script processing and episodic memory network) “models the brain both as a semantic and emotional processor.”
Overloaded with information and competing narratives DISCERN revealed what the researchers called a “hyper-learning hypothesis.” In short, Peck says, in such situations, “the brain assigns an unwarranted importance to new information, (which) can set off delusions and scattered language.”
On first reading the article my automatic reaction was that the article suggested a corollary with the narratives and assumptions political and economic journalists make when confronted by overload in the form of monumental data outputs. Yet the friend who sent the article pointed out the obvious need to distinguish between severe obsessional behaviour and the deductions the person not suffering from mental illness makes.
Nevertheless, he suggested, a posteriori selections – such as how we interpret political developments, stock market noise or opinion poll data – can be a form of mild human delusion.
It is this form of mild human delusion which former Reserve Bank Governor, Ian Macfarlane, was talking about when he observed the way journalists seek to give meaning and construct narratives around discrete pieces of economic information.
So while these mild delusions are not the same as the DISCERN research’s reference to how “an individual may begin to feel that everything is deeply but mysteriously relevant”, they can be a useful explanation of how business, economics and political reporters either miss the signals in all the noise they hear, or just misinterpret the signals and reach erroneous conclusions.
Coincidentally the friend and I were also discussing the work of Stephen Wolfram, a successful business person and controversial scientist, who believes he has discovered a new form of science by looking at how computers codify and replicate human intelligence. Everything, he says, is amenable to computational solutions.
So it might be interesting to see some researchers try to replicate the processes by which some news judgements are made. For instance:
- Creating a narrative to explain unrelated and insignificant facts and isolated events.
- Investing non-events – opinion polls movements within the standard error range – with immense significance.
- Striving for balance by giving as much credence to an MP who thinks gravity is a “theory” and opinions from an architect who thinks Ross Garnaut is a fascist, and that climate change is not real, to scientific facts.
- The infatuation with arguments from authority, for example journalists reporting Twiggy Forrest’s views on the dangers of taxes without mentioning that he has little or no experience of paying them.
If Wolfram is right, a computational solution in which computers codify and replicate the behaviours of journalists, would provide a new way to deconstruct tabloid media print, TV and radio.
Sounds like a great idea for an inter-disciplinary research grant. But one can’t help thinking the explanations might be more mundane.